Finally, after assessing cellular adhesion and the cytoskeletal arrangement, the cellular response was studied via internalization of the complexes through endocytosis. For these investigations, NIH/3T3 fibroblasts were seeded on 25 kDa bPEI-DNA complexes immobilized to PAA-cRGDyK, PAA-RGE, or PAA and chemical inhibitors for three common types of
endocytosis (i.e. macropinocytosis, and clathrin-mediated or caveolae-mediated endocytosis) were added concurrently: amiloride (Amil; 1 mM (278)),
chlorpromazine (CPZ; 10 µg/mL (71)), and genistein (Gen; 200 µM (71)). After culturing the cells for 24 hours, the transfection efficiency was
assessed with flow cytometry and compared to cells seeded without inhibitors on corresponding control substrates (i.e. 25 kDa bPEI-DNA complexes immobilized to PAA-cRGDyK, PAA-RGE, or PAA). The reduction of transfection efficiency in cultured cells was significantly dependent on the culturing substrate (i.e. PAA- cRGDyK, PAA-RGE, or PAA) (Figure 5-7; *; P≤0.05) but, for each culturing substrate, there were no statistical differences relative to the type of inhibitor used. For cells cultured on 25 kDa bPEI-DNA complexes immobilized to PAA- cRGDyK, the fold changes in transfection efficiency comparing cells transfected in the presence of inhibitors to uninhibited cells were 0.19±0.083 for
126 for caveolae-mediated transfection, respectively (Figure 5-7). For cells cultured on 25 kDa bPEI-DNA complexes immobilized to PAA-RGE, the fold changes in transfection efficiency comparing cells transfected in the presence of inhibitors to uninhibited cells were 0.12±0.030 for macropinocytosis, 0.44±0.19 for clathrin- mediated endocytosis, and 0.15±0.050 for caveolae-mediated transfection, respectively (Figure 5-7). Finally, for cells cultured on 25 kDa bPEI-DNA complexes immobilized to PAA, which was the substrate where cultured cells were the least affected by the inhibitors (Figure 5-7), the fold changes in
transfection efficiency comparing cells transfected in the presence of inhibitors to uninhibited cells were 0.65±0.31 for macropinocytosis, 0.87±0.42 for clathrin- mediated endocytosis, and 1.4±0.83 for caveolae-mediated transfection, respectively (Figure 5-7). PA A-c RG Dy K PA A-R GE PA A 0 .0 0 .5 1 .0 1 .5 2 .0 2 .5 F o ld C h a n g e C P Z G e n A m il
Figure 5-7: Endocytic inhibition of NIH/3T3 fibroblasts cultured on 25 kDa bPEI-
DNA complexes immobilized to modified PAA brushes (with cRGDyK and RGE, or no peptides). The transfection of NIH/3T3 fibroblasts was inhibited via the addition of chemicals that modulate endocytosis (i.e. amiloride (Amil; 1 mM), chlorpromazine (CPZ; 10 µg/mL), and genistein (Gen; 200 µM) for
macropinocytosis, and clathrin- and caveolae-mediated endocytosis,
127 tests showed that the culturing substrates (i.e. PAA-cRGDyK, PAA-RGE, PAA) significantly altered the endocytic pathways utilized by the cultured cells
(*:P≤0.05) but there were no statistical differences in the effect of the inhibitors on reducing transfection.
5.4. Discussion
The objective of this work was to tune the ability of PAA brushes to prime the cellular response for SMD transfection through the functionalization of cellular adhesion peptides and the immobilization of different PEI-DNA complexes. After confirming that the PAA brush characteristics (i.e. grafting thickness, swelling behavior) and peptide immobilization (Tables 6-1 and 6-2) were consistent with previously reported results (116, 186, 267) (chapter 3,4), the efficacy of SMD was tested in a model cell line for transfection, NIH/3T3 fibroblasts (132, 150, 251). First, the concentration of the linear GRGDS peptide conjugated to the PAA brushes was investigated, the same peptide used in our previous investigation (267) (chapter 4). Without complexes, substrates conjugated with PAA-GRGDS at 1.5 mg/mL had significantly more cells adhered to the substrate compared to those with PAA-GRGDS at 1.0 mg/mL (and PAA-GRGDS at 0.5 mg/mL; Figure A-1 in Appendix), suggesting that there may be increased transfection in cells cultured on PAA-GRGDS at 1.5 mg/mL. Yet, when fibroblasts were cultured on 25 kDa bPEI-DNA complexes immobilized to PAA brushes conjugated with GRGDS at a higher concentration of 1.5 mg/mL had no significant increase in transfection efficacy in comparison to those on 25 kDa bPEI-DNA complexes immobilized to PAA brushes conjugated with GRGDS at a lower concentration of 1.0 mg/mL. Previous investigations have shown that the RGD density can
128 endothelial cells, fibroblasts) cultured on substrates with higher RGD densities have shown more focal adhesion formation and higher cell spreading and migration rate in comparison to those cultured lower RGD densities (130, 231, 283-286), all of which are cell behaviors that have been shown to improve transfection success (268).
The concentrations of conjugated GRGDS ligands in this work were calculated at a density of 1.0-1.7 µg/cm2, which is within the range of expected values comparable to previously reported functionalized polymer substrates (187, 287-289); yet, close packing of the RGD ligand on the PAA brushes conjugated at a higher concentration may provide insufficient spacing for integrin
engagement in cultured cells (284). Furthermore, according to AFM
measurements, the GRGDS ligands are evenly distributed across the PAA brushes (Figure A-2 in Appendix), and given that the presentation of the RGD ligands in clustered forms has been reported as more effective at promoting cellular adhesion and subsequent transfection (196, 231, 288, 289) rather than evenly distributed RGD ligands, increasing the concentration of GRGDS
conjugated to PAA brushes may not have had an effect on subsequent SMD transfection because of the uniform presentation of the ligand. Therefore, after analyzing the effect of the conjugated GRGDS concentration on transfection in NIH/3T3 fibroblasts, we tested a cyclic RGD ligand (cRGDyK), which has been shown to have protease resistance, high stability, and high affinity for cellular integrins (186, 272), which may improve transfection by upregulating intracellular processing regulated by integrin binding (i.e. focal adhesion formation,
129 endocytosis, and intracellular trafficking). The transgene expression for all
conditions was comparable to our previous investigation (16) (chapter 4) and transfection was highest in cells cultured on 25 kDa bPEI-DNA complexes immobilized to PAA-cRGDyK compared to cells cultured on all other PEI-DNA complexes (i.e. 2 kDa bPEI, 25 kDa LPEI, 2.5 kDa LPEI) immobilized to all other PAA brush substrates (PAA-GRGDS, PAA-RGE, PAA).
After analyzing the effect of the RGD ligand presentation, transfection was assessed with respect to the PEI vector. The MW and branching of PEI vectors are considered the two properties that will dictate the physical properties of PEI- DNA complexes and transfection success (49). In general, branched polymers (i.e. bPEI) and higher MW are considered better transfection vectors because they can more effectively condense the DNA into smaller particles (48, 49); thus, transfection is typically performed using 25 kDa bPEI (46). Yet, in our studies, the size of the complex was not significantly different dependent on the vector in our studies (Figure A-3 in Appendix), which may be due to the salt concentration in the media (i.e. OptiMEM) which can produce larger PEI-DNA complexes (>500 d.nm) that are still able to transfect cells (290, 291)). Given that transfection is typically performed using 25 kDa bPEI (46) and that linear structure and lower molecular weights have been shown to reduce cytotoxicity while maintaining transfection success in cultured cells (252), we examined the four PEI vectors of 2 kDa bPEI, 25 kDa bPEI, 2.5 kDa LPEI, and 25 kDa LPEI to form complexes for immobilization studies and transfection.
130 Although different N/P ratios can be used to optimize transfection with each of these vectors, a N/P ratio of 20 was chosen for our studies to maintain a high level of free PEI in solution to adsorb to the brushes (20). When the total adsorbed mass immobilized to the PAA brushes for complexes formed with 2 kDa bPEI and 25 kDa bPEI was calculated, the value (1.0 µg/cm2) was
comparable to our previous reported values (267) (chapter 3) and the complexes formed with 2.5 kDa LPEI and 25 kDa LPEI had significantly less estimated total absorbed mass on the PAA substrate than those formed with 2 kDa bPEI and 25 kDa bPEI. In our previous paper, we calculated that complexing 0.050 µg of DNA required 0.020 µg of 25 kDa bPEI, suggesting a theoretical total of 0.070 µg of fully formed complexes were adsorbed to the substrate (267) (chapter 4). For the investigations, the amount of DNA immobilized to the substrate was estimated using Cy5-labeled DNA plasmids (Table A-1, Figure A-4 in Appendix), which was 0.050 µg/cm2 for complexes formed with 25 kDa bPEI and 25 kDa LPEI and 0.030 µg/cm2 for complexes formed with 2 kDa bPEI and 2.5 kDa LPEI.
Therefore, we calculated that complexing 0.030 µg of DNA required 0.010 µg of 2 kDa bPEI or 2.5 kDa LPEI, suggesting a theoretical total of 0.040 µg of fully formed complexes were adsorbed to the substrate. Thus, the amount of free PEI immobilized to PAA brushes was estimated as 0.96 µg (1.0-0.040), 0.93 µg (1.0- 0.070), 0.54 µg (0.58-0.040), and 0.55 µg (0.62-0.070) for complexes formed with 2 kDa bPEI, 25 kDa bPEI, 2.5 kDa LPEI, and 25 kDa LPEI, respectively. Thus, complexes formed with LPEI may have had less free PEI adsorbed to the substrate and that may have contributed to the lower transfection success in cells
131 cultured on 2.5 kDa LPEI and 25 kDa LPEI compared to those cultured on
complexes formed with 2 kDa bPEI and 25 kDa bPEI. Given that lower MWs and LPEI is known to produce less transfection in cultured cells (45, 48) and less free polymer may be presented with immobilized complexes formed with 2.5 kDa LPEI and 25 kDa LPEI, the resulting higher levels of transfection in cells cultured on 25 kDa bPEI-DNA complexes (compared to those formed with all other PEI vectors) was consistent with reports that term 25 kDa bPEI as the “gold standard” vector for transfection (46, 292). Moreover, cells cultured on PAA brushes
immobilized with the complexes formed with 2 kDa bPEI and 25 kDa bPEI had slightly higher proliferation compared to cells cultured on PAA brushes
immobilized with the complexes formed with 2.5 kDa LPEI and 25 kDa LPEI. The slight increase in proliferation of cells cultured on complexes formed with 2 kDa bPEI and 25 kDa bPEI may have contributed to the increase in transfection success, as proliferation is commonly associated with successful internalization and nuclear entry due to the compromised integrity of the nucleus in dividing cells (132). The results of the proliferation assay that immobilizing the PEI-DNA complexes (and adsorbed free PEI) to PAA brushes does not cause cytotoxicity (possibly even improving proliferation as shown in cells cultured on bPEI-DNA complexes at both MWs) is exceptional, as vector cytotoxicity in cells cultured on TCPS and other substrates has often been cited as a significant barrier to
transfection success with PEI-DNA complexes (56, 293, 294).
Given the results for transfection and proliferation, we further investigated the cellular response of cells cultured on 25 kDa bPEI-DNA complexes
132 immobilized to PAA-cRGDyK in comparison to cells cultured on 25 kDa bPEI- DNA complexes immobilized to control PAA-RGE and PAA. First, the cell density showed that cells cultured on 25 kDa bPEI-DNA complexes immobilized to PAA- cRGDyK had a similar cell count for each image and the most focal adhesions per cell compared to cells cultured on PAA-RGE and PAA, which indicates that the cells were more adhered to the substrates, presumably through integrin binding to the RGD ligand (187, 273). In contrast, cells on 25 kDa bPEI-DNA complexes immobilized to PAA-RGE had a wide distribution of cell densities from 1 cell per image to 60 cells per image, which suggested there were islands of confluent cells and areas of empty culture space on these substrates (i.e. PAA- RGE) rather than evenly distributed cultured cells. Healthy growth of fibroblasts is typically in an even monolayer (295), and cellular aggregation is commonly
mediated by cell-cell adhesion that can increase stress fiber formation (296), which agreed with our images that showed cells cultured on 25 kDa bPEI-DNA complexes immobilized to PAA-RGE had vinculin staining on cells that appeared to be adhered to one another rather than the substrate or focal adhesions and these cells also had the most actin stress fibers per cell. Thus, given that vinculin can also mark adheren junctions for intercellular adhesion (297) and that stress fibers are also known to stabilize protein complexes at those junctions (298), transfection in cells cultured on 25 kDa bPEI-DNA complexes immobilized to PAA-RGE may have been stimulated through the intercellular interactions and intracellular trafficking along the stress fibers (125), rather than cell-material interactions. Finally, actin staining also showed that the morphology of cells that
133 were cultured on 25 kDa bPEI-DNA complexes immobilized to PAA were
flattened and spread but did not have the structural organization of cells cultured on PAA-cRGDyK and PAA-RGE, suggesting that there was less cell adhesion to PAA brushes (226), as expected, but the immobilized complexes (and free PEI) still enable cell adhesion to occur on PAA brushes that are typically cell-repellent without peptides (116, 153) (chapter 3).
Along with the results for the cellular response, DAPI stain for the nucleus also stained the DNA plasmids in the formed 25 kDa bPEI-DNA complexes, as it binds strongly to adenine-thymine rich regions in DNA (299). In images of the stained cells, the plasmid DNA of the complexes is especially apparent in the substrates with PAA alone (i.e. no peptide), suggesting that there is low internalization of the complexes from cells cultured on PAA alone, which may contribute to their low levels of transfection. Furthermore, the conclusion that there is significantly less internalization of 25 kDa bPEI-DNA complexes into cells cultured on PAA compared to those cultured on PAA-cRGDyK is supported by the results of inhibiting endocytosis, which had a minimal effect on transfection efficiency in cells cultured on PAA in comparison to those cultured on PAA- cRGDyK and PAA-RGE.
Along with the low effect of the inhibitors on cells cultured on PAA alone, investigations into the endocytic pathways showed that the effect of the inhibitors was more significant in cells cultured on substrates with peptides (i.e. PAA- cRGDyK, PAA-RGE), as RGD peptides are known to enhance endocytosis and transfection (73, 108, 111, 125, 196, 274, 275). Chlorpromazine (which causes
134 clathrin to localize and accumulate in late endosomes, thereby preventing
endosomal escape of complexes (300)) was the least effective at inhibiting
transfection efficiency, suggesting that clathrin-mediated endocytosis was not the most efficient pathway for transfection in our system, in agreement with previous reports that clathrin-mediated endocytosis is optimal for lipid-based transfection rather than polymers such as PEI (71, 300-302). Moreover, genistein, which prevents vesicle formation in caveolae-mediated endocytosis (303), had more success at reducing transfection efficiency than chlorpromazine, which is supported by previous investigations that have cited caveolae-mediated
endocytosis as a more efficient endocytic pathway for transfection (compared to clathrin-mediated endocytosis) (77), especially with polyplexes in both bolus (109, 300, 304) and substrate-mediated (60) delivery formats. Therefore, the investigations into the endocytic mechanisms suggest that the PEI vector (free and complexed) influenced the internalization pathway in tandem with the cRGDyK ligand.
Overall, the most effective inhibitor for cells cultured on 25 kDa bPEI-DNA complexes immobilized to substrates (i.e. PAA-cRGDyK, PAA-RGE, PAA) was amiloride. Amiloride has been cited as effective at decreasing macropinocytosis by lowering submembranous pH and preventing signaling from the RhoGTPases Cdc42 and Rac1 (278), which are known for contributing to focal adhesion
formation (111, 128, 305). Macropinocytosis, clathrin-, and caveolae-mediated endocytosis\ have all been shown to be modulated by focal adhesion formation (305, 306). Thus, substrates that promote focal adhesion formation in cultured
135 cells (i.e. PAA-cRGDyK) may increase transfection success in those cultured cells through increased endocytosis via all three common pathways, but
especially via macropinocytosis. Finally, another reason that macropinocytosis may have been the most optimal pathway for transfection may simply be related to the size of the complexes (i.e. ~600 nm, Figure A-3 in Appendix), as receptor- mediated pathways such as clathrin- and caveolae-mediated endocytosis
typically takes up complexes under 200 nm in diameter (72). Thus, although the results of the investigations into the cellular response (i.e. proliferation, focal adhesion formation, and endocytosis) strongly suggested that cells can be primed by the substrate to increase transfection success, factors such as the characteristics of the complex (i.e. size, charge) are important to consider in investigations to improve transfection.
5.5. Conclusions
In our previous studies, we showed that PAA brushes can be “grafted-to” Ti substrates and RGD can be conjugated to these brushes to support cell
adhesion, and those PAA-RGD modified Ti substrates can be used as a platform to immobilize 25 kDa bPEI-DNA complexes to transfect NIH/3T3 fibroblasts via SMD. Given that the presence of the RGD ligand and the presence of the free PEI may have synergistically contributed to enhanced transfection via SMD in fibroblasts cultured on 25 kDa bPEI-DNA complexes immobilized to PAA- GRGDS substrates compared to transfection in cells cultured on substrates without the RGD ligand and without free PEI, herein we investigated tuning these
136 factors through the concentration and binding affinity of the RGD ligand (i.e. cyclic vs. linear) and the branching and the MW of the PEI to prime the cellular response to transfection. After determining the optimal priming conditions (complexes formed with 25 kDa bPEI immobilized to PAA brushes conjugated with cRGDyK), the cellular response was investigated. Increased proliferation of cells cultured on 25 kDa bPEI-DNA complexes immobilized to PAA brushes may have increased the nuclear availability to the complexes, which may have
contributed to the transfection success. Furthermore, the presentation of 25 kDa bPEI-DNA complexes and free bPEI adsorbed to the PAA brushes may have mitigated the cytotoxicity effect of culturing cells on PEI-DNA complexes. Further investigations into the cellular response show that cells cultured on PAA-
cRGDyK had increased focal adhesion formation, presumably related to integrin binding of the RGD ligand, which may have led to increased endocytosis of the complexes (especially via macropinocytosis), although further tuning of the RGD density and presentation (i.e. clustering) may enhance the improvement in transfection. Overall, the findings of this chapter suggest that the modification of Ti with PAA brushes is a tunable method to affect the efficacy of nonviral gene delivery, that there is a synergistic effect of free PEI and the RGD ligand on the cellular response to transfection, and that PAA-cRGDyK may have future applications to modify substrates that could be improved by gene delivery including biomedical devices, implantable sensors, and diagnostics tools.
137
CHAPTER 6 Conclusions and Future Directions
6.1. Conclusions
Cell-biomaterial interactions that occur on a substrate can modulate the cellular processes related to successful nonviral gene delivery, such as adhesion and proliferation (307-309), migration (310, 311), and endocytosis (312, 313). Along with priming cellular responsiveness to gene delivery, biomaterial interfaces can be used to immobilize formed DNA complexes through electrostatic interactions or covalent binding in a process termed “substrate- mediated gene delivery” or SMD (19, 20, 59) to enhance transfection by presenting DNA within the microenvironment of the cell. Nonviral SMD
investigations have not previously been performed on clinically relevant metallic biomaterials (e.g. titanium (Ti) (2)), but Ti is used ubiquitously in medical devices and implants whose integration and functionality could be further improved with gene delivery as shown in a previous viral investigation into SMD on Ti (314) Given that nonviral gene delivery is safer but less efficient compared to viral vectors, there is a need for a cell-material interface that modulates the cellular response and immobilization of nonviral DNA complexes onto Ti biomedical implants and devices. Thus, a novel platform for SMD (chapters 3, 4, and 5) was investigated by chemically altering the cell-material interface through grafting of stimuli-responsive poly(acrylic acid) brushes (PAA) to Ti, and conjugating the PAA brushes with arginyl-glycyl-aspartic acid (RGD) ligands, which showed enhanced transfection facilitated by the cellular response to the interface, as well
138 as the ability of the brushes to sequester adjuvant-like free PEI. These
substrates may immobilize therapeutic DNA complexes for applications such as Ti biomedical devices, implantable sensors, and diagnostics tools.
In chapter 3 of this dissertation, we described the development and characterization of a simple method of grafting PAA brushes to Ti substrates, which amplified the substrate functionality through the high density of COOH groups that deprotonate in response to pH-stimuli and allow for the ability to